click any column header to sort

TTS Bench — Speed — mac-default

Rig: mac-m4 — Apple M4 (10C) · Apple M4 GPU (MPS) · 16 GB RAM · Darwin 25.5.0
Label: default voice
TTFA = time to first audio (ms; lower is better). RTF = real-time factor (× realtime; higher is better; e.g. 10× means 10 sec of audio generated per 1 sec of compute). Cold = first run after process start; warm = subsequent runs.

Speed winners

Fastest predefined-voice: Piper (cpu) — 32.14× warm RTF, 208ms warm TTFA

Fastest cloning-capable: Pocket-TTS (cpu) — 7.82× warm RTF, 44ms warm TTFA

ModelDeviceTTFA coldTTFA warmRTF coldRTF warmPeak RAMPeak VRAMSizeReleased
Pipercpu228ms208ms27.47×32.14×561 MB~25MBJan 2023
Kokoromps920ms489ms7.58×15.28×1.26 GB82MDec 2024
MeloTTSmps1.88s512ms4.09×15.13×1.55 GB~52MFeb 2024
Kokoro (MLX) (4/5 ok)mps2.10s514ms3.90×14.88×774 MB82M
Kokorocpu914ms749ms7.78×10.04×2.49 GB82MDec 2024
KittenTTS Nano 0.1cpu994ms1.01s8.10×8.08×385 MB<100MAug 2025
Pocket-TTScpu89ms44ms7.50×7.82×1.69 GB100MJan 2026
Soprano 1.1 80Mcpu912ms906ms7.15×7.15×1.18 GB80MJan 2026
MeloTTScpu2.35s1.27s3.20×6.48×3.30 GB~52MFeb 2024
Soprano 1.1 80Mmps2.48s1.30s2.76×5.16×1.64 GB80MJan 2026
Supertonic 3cpu1.86s1.57s4.18×4.81×968 MB99MMay 2026
StyleTTS 2cpu5.94s2.21s2.30×3.82×3.13 GB~148MJun 2023
NeuTTS Nanocpu740ms266ms2.30×3.10×5.83 GB229MDec 2025
NeuTTS Nanomps1.45s464ms1.81×2.96×1.28 GB229MDec 2025
NeuTTS Aircpu1.38s358ms1.55×2.16×6.12 GB748MSep 2025
NeuTTS Airmps2.32s583ms1.30×2.08×1.55 GB748MSep 2025
Coqui XTTS-v2mps17.81s7.82s1.14×1.70×1.84 GB750MOct 2023
Chatterbox Turbomps9.95s5.07s0.75×1.44×2.30 GB744MDec 2025
Coqui XTTS-v2cpu6.01s5.48s1.41×1.43×4.61 GB750MOct 2023
Voxtral 4B TTSmps6.45s6.37s1.12×1.19×1.93 GB4BNov 2025
Chatterbox Turbocpu6.53s6.45s1.07×1.14×4.80 GB744MDec 2025
VibeVoice Realtime 0.5Bmps9.56s8.30s0.98×1.13×838 MB0.5BDec 2025
OmniVoicemps8.05s12.20s0.81×0.79×1.38 GB~1BMar 2026
OmniVoicecpu12.84s11.28s0.49×0.57×3.59 GB~1BMar 2026
Magpie-TTScpu25.55s26.19s0.40×0.40×4.69 GB357MDec 2025
Parler-TTS Mini v1cpu25.23s22.58s0.32×0.39×8.60 GB878MJun 2024
Chatterboxcpu18.77s17.08s0.33×0.37×4.83 GB1.2BApr 2025
VibeVoice Realtime 0.5Bcpu25.45s25.10s0.38×0.36×3.17 GB0.5BDec 2025
Chatterboxmps25.95s31.35s0.27×0.29×698 MB1.2BApr 2025
Qwen3-TTS 1.7B Basecpu24.52s25.11s0.25×0.24×3.93 GB1.7BJan 2026
Sesame CSM-1Bcpu38.67s31.30s0.19×0.24×6.08 GB1BMar 2025
VoxCPM2 2Bcpu28.36s52.13s0.29×0.17×144 MB2BApr 2026
IndexTTS-2cpu80.00s48.40s0.08×0.14×318 MB1.5BJun 2025
F5-TTS v1mps45.43s45.54s0.09×0.09×1.29 GB330MOct 2024
F5-TTS v1cpu47.26s47.15s0.09×0.09×3.31 GB330MOct 2024
VibeVoice 1.5B (2/4 ok)cpu71.46s70.84s0.05×0.06×9.01 GB1.5BAug 2025
OuteTTS 1.0 1B (2/5 ok)mps121.18s93.13s0.04×0.05×733 MB1BApr 2025
OuteTTS 1.0 1B (2/5 ok)cpu100.71s95.88s0.05×0.05×7.38 GB1BApr 2025
LuxTTScpuLuxTTS install failed (piper-phonemize has no Windows wheels)
LuxTTSmpsLuxTTS install failed (piper-phonemize has no Windows wheels)
ZipVoice 123McpuReference wav missing for voice cloning
ZipVoice 123MmpsReference wav missing for voice cloning
Mars5-TTScpuTimed out after 10 min — model too slow at this prompt length
VibeVoice 1.5BmpsSkipped — out of memory (model exceeds available RAM)